To the Editor: Our study (tumour necrosis factor alpha levels in hepatitis B virus related chronic active hepatitis and liver cirrhosis and its relationship to Knodell and Child–Pugh scores) was aimed to investigate the possibility that circulating tumour necrosis factor alpha (TNF-α) levels might represent an effective indicator of the disease in patients with chronic active hepatitis B (CAH-B) and liver cirrhosis (LC). Plasma samples of 30 subjects from each group were taken and analysed using enzyme-linked immunosorbent assay technique. Results were compared with those of 30 healthy controls. As a result of our study, we concluded that plasma TNF-α levels are significantly elevated in patients with CAH-B proportionally with hepatitis activity index. This data is in agreement with those of previous studies. On the other hand, we found out an inverse relationship between Child–Pugh score and plasma TNF-α level in LC patients. This later result is in contradiction with the previously reported data as mentioned in the original paper. Tumour necrosis factor is a pleiotropic proinflammatory cytokine, mainly produced by activated macrophages and in smaller amounts by several other types of cells, with strong antitumoural capacities both in vitro and in vivo. TNF-α plays a main role in the extrinsic pathway of apoptosis. So, it serves the clearance of virus infected hepatocytes (1). As a matter of fact, TNF-α and other proinflammatory cytokines were reported to be elevated parallelly with transaminase levels (2,3). There are several reasons for elevated TNF-α level in cirrhotic patients. These are mostly related to concomitant factors such as exposure to bacterial lipopolysaccharids generated from apparent or unapparent bacterial infections, existence of hepatic encephalopathy or other complications associated with advanced cirrhosis or even poor psychological status (4–6). On the other hand, there are also a number of factors which may negatively influence the level of TNF in cirrhotic patients. In addition to reduced amounts of endogenous cells (Kupffer cells, endothelial cells, hepatocytes and bile duct epithelial cells), quantitative and functional abnormalities in other cellular source of TNF-α production such as natural killer cells has also been reported in cirrhosis (7–9). Furthermore, adipose tissue which is capable of TNF-α release, is commonly decreases in advanced cirrhosis (10). Similarly, peripheral blood neutrophilic granulocytes can also produce and release TNF-α (11). Hypersplenism because of the enlarged spleen as a result of portal hypertension may reduce neutrophil count and consequently TNF-α level. Increased plasma volume and dietary factors may also contribute to the alteration of the plasma level of TNF-α. We have carefully searched the presence of clinical or laboratory evidence of concomitant infectious and inflammatory diseases in our cases. All the cases were clinically stable. Finally, we believe that marked increase of TNF-α production in advanced cirrhosis is possibly determined by concomitant infections and several complications. Further studies on larger series of highly selected cases are necessary to confirm this hypothesis. Tumour necrosis factor plays an important role in host defence and tumour growth control. Therefore, its deficiency or anti-TNF antibody therapies may increase the risk of serious infections and malignancies (12,13). If reduction of TNF-α production in advanced cirrhosis is established, it will possibly be considered as one of the predisposing factors for increased risk of infection and carcinogenesis.